Transition of the tumor microenvironment with clonal evolution of hepatocellular carcinoma.

Abstract

467 Background: The combination of atezolizumab and bevacizumab has become a standard treatment for advanced hepatocellular carcinoma (HCC). The indications for combination immunotherapy, including adjuvant therapy and combination with local therapy, are expected to expand in the future. Although exploring the tumor microenvironment has essential clinical implications in the era of combination immunotherapy, most studies on the tumor microenvironment in HCC are based on the analysis of archival samples at the time of HCC diagnosis. Considering the lengthy clinical course and characteristics of multicentric carcinogenesis, the tumor microenvironment may not be constant, but may differ from that at the time of advanced HCC. The present study aimed to assess the changes in the tumor microenvironment during the evolution to advanced HCC using archival samples and samples obtained prior to systemic therapy from the same patients. Methods: The tumor microenvironment was compared in 20 cases by immunohistochemical analyses of CD8, PD-L1, and VEGF expression. Levels of CD8-positive lymphocytes were defined as the mean number of CD8-positive lymphocytes in the tumor per 1 mm2 and classified as low or high infiltration using the median value. PD-L1 expression was classified as negative (< 1%) or positive (≥1%). VEGF expression was classified as low (< 50%) or high (≥50%). Of the 20 cases, gene expression analysis was performed by the PanCancer IO360 Panel in 14 cases. Results: Approximately 40% of the archived samples were obtained at the early stage. Comparison of the tumor microenvironment between the two time points showed that 35% of the patients changed to CD8-positive lymphocyte high infiltration, 20% to positive PD-L1 expression, and 29% to high VEGF expression. The result of comparing gene profile was different between the archival samples and those prior to systemic therapy in 64% of the patients on the expression heat map. Gene set enrichment analysis showed significant changes in T cell cytotoxicity between the two time points. Conclusions: The tumor microenvironment might not be constant, but may change during the evolution to advanced HCC. In the present study, based on the results of the gene mutation analysis using whole-exome sequencing, we also reported the analyses by carcinogenic patterns such as multicentric occurrence and intrahepatic metastasis.